How to Explain a Traffic Drop to Your Boss or Client Without Overpromising

Nobody writes about this part, and it decides more recovery projects than the technical work does. A correct diagnosis explained badly loses budget. A vague diagnosis explained confidently keeps it, right up until the month the numbers do not move and the credibility goes all at once.

What follows is how I structure these conversations, and the specific things I have learned not to say.

Lead with the number and the date

Not the cause, not the plan. The two facts that establish you actually know what happened.

Something like: organic traffic fell 43 percent starting 12 March, measured against the 28 days before that date. Precise, verifiable, and it immediately tells the person listening that this has been measured rather than noticed.

Define the baseline explicitly, because the alternative is that somebody else defines it later and picks your best month. A twenty-eight day average from a normal period is defensible. A peak is not, and you will spend the rest of the project measured against an outlier.

Separate what you know from what you suspect

This is the discipline that protects you three months later, and it costs nothing at the time.

Say this	Not this	Why it matters
The drop began 12 March and correlates with a confirmed core update	The March core update hit us	Correlation is what you have. Causation is what you are testing
The loss is concentrated on our guide pages	Google devalued our content	One is observed, the other is an interpretation
Reassessment happens at the next core update, historically 7 to 147 days apart	We should see recovery in about three months	You do not control the date and cannot be held to it
Full recovery happens in roughly three cases in ten	We will get the traffic back	Sets an expectation you can actually meet

The third row is the one people resist hardest, because a range that wide sounds like you do not know. It is more useful to say it in week one than to be asked in month four why the three-month estimate passed without anything happening.

The update template I use

Short enough to be read, structured enough that the same fields appear every time so progress is visible across updates rather than re-argued in each one.

• Status: one line. Diagnosing, implementing, or awaiting reassessment.
• Current position: traffic against the stated baseline, as a percentage, with the baseline restated every time.
• What we know: confirmed facts only.
• Working hypothesis: labelled as such.
• Completed since last update: specific, with dates.
• Next reassessment: what has to happen before results are possible, and the honest range.
• What would change the plan: the thing you are watching for.

The last field is the one that earns trust. Saying in advance what evidence would make you change course is how you demonstrate you are testing a theory rather than defending one.

Explaining the wait

The hardest thing to convey to someone outside SEO is that work completed in February can sit unassessed until May, and that this is normal rather than a failure.

The framing that lands best in my experience is an exam rather than a machine. Core ranking systems re-score sites during core updates, not continuously. You can revise perfectly and still have to wait for the paper to be marked, and nobody publishes the exam timetable.

Concrete numbers help more than analogies though. Between the November and December 2024 core updates there were seven days. Between June and December 2025 there were 147. Quoting the 82-day average without that range is how a project ends up feeling like a failure at month four.

What not to say

• Google penalised us. Unless there is a manual action in Search Console, there is no penalty. The word sets an expectation of an appeal that does not exist.
• We should recover by a given date. You do not control the reassessment. Give the range and the mechanism instead.
• Everyone in the industry got hit. Sometimes true and always sounds like deflection. If competitors held, it is false and somebody will check.
• It is just an algorithm change, nothing we can do. Abdication. Core updates reassess what is there now, so there is always something to change.
• Silence until there is good news. The most common mistake. Absence of updates reads as absence of work.

If you have to say recovery may not be full

Say it early. It is a reasonable thing to explain in week two and a very difficult thing to introduce in month six.

The reasons are concrete and easy to state: competitors improved while we were down, some of our queries are searched less than they were, and the results page now gives fewer clicks to the same position than it did two years ago. None of that is an excuse, and all of it is checkable. The detail is in my recovery stalled at partial.

Where the honest answer is that recovery is not worth the spend, that conversation is better had with a calculation than with a feeling. The traffic loss calculator gives you a figure to put in front of people.

Frequently asked questions

How do I explain a traffic drop to my boss?

Lead with the date it began and the percentage lost against a stated baseline, then separate confirmed facts from your working hypothesis, then give the reassessment mechanism and its honest range rather than a recovery date.

Should I promise a recovery date?

No. Core update timing is not under anyone’s control and intervals have ranged from seven to one hundred and forty-seven days. Give the range and explain what has to happen before results are possible.

How often should I update stakeholders during a recovery?

Regularly enough that silence never becomes the message, with the same fields each time so progress is visible across updates. Fortnightly works for most projects, and a short update with no movement is better than no update.

Should I say the word penalty?

Only if there is a manual action in Search Console. Otherwise it is a demotion, and calling it a penalty creates an expectation of an appeal process that does not exist for algorithmic changes.